Since the discovery of fetal hemoglobin, augmentation of its synthesis has been proposed as a therapeutic goal in patients with beta thalassemia. Unfortunately, clinical trials of most pharmacologic agents have demonstrated modest clinical responses only. Guided by previous work, the primary hypotheses of this research proposal is that augmentation of fetal hemoglobin, sufficient to increase steady-state total hemoglobin concentration and reduce transfusion requirements, will be most effective during sequential administration of selected agents to patients with specific mutations within the beta globin cluster and that responses will also be influenced by the degree of erythroid marrow expansion, with the most marked responses observed in patients in whom marrow expansion is maximized before therapy. The specific aim is to determine increases in steady-state total and fetal hemoglobin concentrations, reduction in globin chain imbalance, toxicity, and patient compliance associated with selected pharmacologic regimens with patients with genotypes selected to test the above hypotheses. To achieve this aim, 30 patients with three groups of mutations, including specific deletions or rearrangements of putatively important regulatory elements within the beta globin cluster, will be recruited. In the first study phase, patients will be treated with sodium phenylbutyrate, hydroxyurea and recombinant erythropoietin alone and in combination. In the second study phase, low-dose subcutaneous cytosine arabinoside will be offered to adult patients. The key study comparison will be made with respect to the total hemoglobin concentration; we estimate that the minimal clinically important difference between the hemoglobin concentration at baseline and after treatment will be 3 grams/deciliter. Compliance with oral therapy will be monitored using computerized pill containers, and with subcutaneous therapy using counts for syringes returned each visit. This study represents an attempt to provide definitive therapy for the primary abnormally in severe beta thalassemia, severe imbalance of globin chains, and to reduce or eliminate transfusions in selected patients.